On the scalability of hypergraph models for sparse matrix partitioning

Abstract : We investigate the scalability of the hypergraph-based sparse matrix partitioning methods with respect to the increasing sizes of matrices and number of nonzeros. We propose a method to rowwise partition the matrices that correspond to the discretization of two-dimensional domains with the five-point stencil. The proposed method obtains perfect load balance and achieves very good total communication volume. We investigate the behaviour of the hypergraph-based rowwise partitioning method with respect to the proposed method, in an attempt to understand how scalable the former method is. In another set of experiments, we work on general sparse matrices under different scenarios to understand the scalability of various hypergraph-based one- and two-dimensional matrix partitioning methods.
Complete list of metadatas

Cited literature [9 references]  Display  Hide  Download

https://hal-ens-lyon.archives-ouvertes.fr/ensl-00529088
Contributor : Bora Uçar <>
Submitted on : Sunday, October 24, 2010 - 7:28:44 PM
Last modification on : Wednesday, November 20, 2019 - 3:18:39 AM
Long-term archiving on : Friday, October 26, 2012 - 12:15:29 PM

File

ucca10.pdf
Files produced by the author(s)

Identifiers

Collections

Citation

Bora Uçar, Umit Catalyurek. On the scalability of hypergraph models for sparse matrix partitioning. 18th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP 2010), IEEE, Feb 2010, Pisa, Italy. pp.593-600, ⟨10.1109/PDP.2010.92⟩. ⟨ensl-00529088⟩

Share

Metrics

Record views

458

Files downloads

290